Stop Copy-Pasting Legacy Parts into AM Tools

One of the most persistent habits in additive manufacturing is also one of the most limiting. Engineers take an existing CAD model (usually designed for machining, casting, or any other traditional manufacturing technique) import it into a DfAM workflow, and then apply topology optimisation or latticing  in the hope of “unlocking” AM’s potential. 

The output often looks impressive. Organic geometries. Lightweight structures. Intricate lattices. But beneath the surface, the underlying design logic has rarely changed. In many cases, the part is still solving the wrong problem. 

The Hidden Constraint in Modern DfAM 

When a design originates from traditional manufacturing assumptions (such as tool access, draft angles, or assembly constraints) those assumptions shape the architecture of the component itself. By the time  these “standardised” DfAM  workflows are applied, the design space has already been constrained. 

What follows is DfAM within inherited boundaries,  without questioning whether the starting geometry was appropriate in the first place. 

The result is a familiar pattern in additive manufacturing. Parts that appear radically different but still reflect legacy design thinking. 

This is why many “optimised” AM components look surprisingly similar. They are variations on a theme defined by software presets and historical constraints rather than engineering intent. 

The Difference Between Adaptation and Design 

At Metamorphic AM, we see this distinction as central to the future of DfAM. Adaptation asks “How can this existing part be modified to work in additive manufacturing?” Engineering asks a different question entirely, “What is the best geometry to achieve the required performance?” Only once that question is answered should computational tools enter the process. 

This approach transforms design for additive manufacturing from a post-processing exercise into a true design discipline. 

Instead of adapting legacy geometry, we begin with the functional requirements of the component (structural loads, thermal behaviour, flow characteristics, vibration response, manufacturability constraints, and scaling considerations). 

From there, computational tools — including parametric modelling, simulation, and multi-objective exploration — help map the design space. But they are used to inform engineering decisions, not replace them.  

Geometry as a Record of Intent 

When DfAM is approached from first principles, geometry stops being decorative. It becomes purposeful. 

A braided structure might exist because it improves heat transfer or fluid mixing. A novel lattice structure may be programmed to deform in a highly controlled manner during thermal expansion. Complex textures may be added to a component’s surface to minimise eddy currents. 

In other words, complexity becomes a consequence of performance. 

This philosophy is central to the way Metamorphic approaches computational design. Our workflows combine engineering judgement with simulation and custom parametric tools to explore trade-offs between performance, manufacturability, and scale. 

Because in the real world, no component optimises a single objective. Every design is a balance. 

The Real Opportunity in Additive Manufacturing 

Additive manufacturing offers something extraordinary, and that is the freedom to rethink geometry itself. But that opportunity is lost if the design process simply starts with legacy parts and attempts to optimise them. 

The true promise of AM lies not in modifying existing components, but in re-imagining them. That means questioning assumptions, redefining design boundaries, and treating computational tools as collaborators rather than decision-makers. 

It also means recognising that preset DfAM tools don’t create innovation. Good engineering does. As additive manufacturing matures, the companies that succeed will not be those who generate the most visually complex geometries. They will be the ones who ask better design questions from the beginning. 

Because the future of DfAM will not be defined by software. It will be defined by engineering intent.